Agentic RAG là gì ?

📹 VIDEO TITLE 📹

What is Agentic RAG ?

✍️VIDEO DESCRIPTION ✍️

In this video, we start by revisiting Retrieval-Augmented Generation (RAG), a powerful technique that enhances language models by enabling them to retrieve external information before generating a response. RAG bridges the gap between static knowledge embedded in a model and dynamic or domain-specific information stored in external sources like vector databases. However, traditional RAG pipelines operate in a fixed, single-step retrieve-and-generate loop — limiting their ability to handle more nuanced, multi-step tasks.

Next, we explore ReAct-style agentic workflows, where an AI agent can reason step-by-step and take actions — like calling tools or issuing new queries — based on intermediate observations. These agentic workflows introduce autonomy and adaptability into the system, enabling the model to break down problems, revise its plan, and iterate toward a solution. By combining reasoning and action, agents can better tackle complex, ambiguous, or evolving queries that go beyond what a single pass can solve.

Finally, we bring these two paradigms together to introduce Agentic RAG, a next-generation architecture that fuses RAG's retrieval power with the dynamic reasoning of agents. In Agentic RAG, retrieval becomes a tool the agent can call repeatedly, using planning, reflection, and tool use to iteratively improve results. This design pattern unlocks more accurate, complete, and intelligent systems — ideal for research tasks, multi-hop QA, and any use case requiring thoughtful, tool-augmented generation.

🧑‍💻GITHUB URL 🧑‍💻

No code samples for this video

📽OTHER NEW MACHINA VIDEOS REFERENCED IN THIS VIDEO 📽

Build an MP Neuron with PyTorch - https://youtu.be/L6FrRQEe3GY

LangChain versus LangGraph - https://youtu.be/JaCSgQtziMA

Chroma versus Pinecone Vector Database - https://youtu.be/EtR6BWrCbMQ

What is the Chroma Vector Database ? - https://youtu.be/qn738hVKJe4

RAG with OpenAI & Pinecone Vector Database ? - https://youtu.be/IuXVTJm-iF8

What are LLM Function Calls ? - https://youtu.be/Nh6qoBnreBc

Embeddings with Open AI & Pinecone Vector Database - https://youtu.be/GgeoyzWBrSI

What is Hugging Face? - https://youtu.be/QvO4EnN905Y

RAG vs Fine-Tuning - https://youtu.be/AJmlg7rdmLA

What is RAG ? - https://youtu.be/SDsY9hHS9Qo

What is the Perceptron? - https://youtu.be/UeKxO-Sk0wE

What is the MP Neuron? - https://youtu.be/MBSHhsvaTjs

What is Physical AI ? - https://youtu.be/Xya21TpCog0

What is the Turing Test ? - https://youtu.be/wXMLF54AUwU

What is LLM Alignment ? - https://youtu.be/nYX73hSDEqo

What are Agentic Workflows? - https://youtu.be/CwLAtLyFiTM

Why is AI going Nuclear? - https://youtu.be/eFYy1UYzdZg

What is Synthetic Data? - https://youtu.be/34n9DxFqFc0

What is NLP? - https://youtu.be/C528qW0Zr8k

What is Open Router? - https://youtu.be/pfT6l0yMsB0

What is Sentiment Analysis? - https://youtu.be/hkmAuBWhiXs

What is Mojo ? - https://youtu.be/5uqEPn3DQl8

SDK(s) in Pinecone Vector DB - https://youtu.be/ttnPUbiLjv0

Pinecone Vector DB POD(s) vs Serverless - https://youtu.be/t7qpxjTTccc

Meta Data Filters in Pinecone Vector DB - https://youtu.be/ztXrf88sX-M

Namespaces in Pinecone Vector DB - https://youtu.be/ztXrf88sX-M

Fetches & Queries in Pinecone Vector DB - https://youtu.be/ztXrf88sX-M

Upserts & Deletes in Pinecone Vector DB - https://youtu.be/ztXrf88sX-M

What is a Pineconde Index - https://youtu.be/IHm0-WBELTI

What is the Pinecone Vector DB - https://youtu.be/IHm0-WBELTI

What is LLM LangGraph ? - https://youtu.be/w4U3gG_C4VY

AWS Lambda + Anthropic Claude - https://youtu.be/WaxYMhNsCAk

What is Llama Index ? - https://youtu.be/vz3Z2XETpGM

LangChain HelloWorld with Open GPT 3.5 - https://youtu.be/tD335RLNYJQ

Forget about LLMs What About SLMs - https://youtu.be/Pn7a35dQq2s

What are LLM Presence and Frequency Penalties? - https://youtu.be/J66CRz6s734

What are LLM Hallucinations ? - https://youtu.be/4xmMj6UPIb4

Can LLMs Reason over Large Inputs ? - https://youtu.be/nCVjjXPIrxc

What is the LLM’s Context Window? - https://youtu.be/y5wBbDSe0cM

What is LLM Chain of Thought Prompting? - https://youtu.be/Lwn88e17u4k

Algorithms for Search Similarity - https://youtu.be/jaJd9IFlVCA

How LLMs use Vector Databases - https://youtu.be/1GT6ctTyXFo

What are LLM Embeddings ? - https://youtu.be/UShw_1NbpCw

How LLM’s are Driven by Vectors - https://youtu.be/Yl_ebS_jWZM

What is 0, 1, and Few Shot LLM Prompting ? - https://youtu.be/ckQPDM-97dM

What are the LLM’s Top-P and TopK ? - https://youtu.be/aDmp2Uim0zQ

What is the LLM’s Temperature ? - https://youtu.be/_YTnZOYxSjE

What is LLM Prompt Engineering ? - https://youtu.be/s_8Ba_UJkcA

What is LLM Tokenization? - https://youtu.be/q77s1gurXYU

What is the LangChain Framework? - https://youtu.be/dS5H-bjItqE

🔠KEYWORDS 🔠

#AgenticRAG

#RetrievalAugmentedGeneration

#RAG

#AgenticWorkflows

#ReAct

#LLM

#VectorDatabases

#AIworkflow

#LLMTools

#SemanticSearch